Today’s Best Band Ever™ is Fontaines DC. They’re one of the brightest new bands out there.
So how’s the world?
People glue themselves to famous artwork to raise awareness to the ecology. Noble and stupid. But hey, if humanity is OK with killing their only home, maybe superglue is the solution.
It’s the 9th year of a major continuos war in Europe.
A rich clown has bought a popular website.
Business as usual.
Among many other categorisations, I’d like to point out two kinds of cloud customers - profit centres and cost centres.
In the cloud context, profit centres are those that make money directly by using computers. When they go to the cloud, they’re interested in making more money and faster.
Cost centres are departments in the traditional companies that have to use computers, but the profits come from elsewhere. When these companies go to the cloud, their objective is to spend less money.
Of course, there are mixed cases, where a traditional company starts to realise that the software is slowly eating the world, they declare that they’re now a digital company, ask a big consultancy to create a cloud strategy for them and, ironically, task the cost centre with the implementation of it. Because “they’re doing IT anyway”.
There’s the opposite case, where a fast moving digital company suddenly realises that they might actually need to finally address cybersecurity and compliance topics. Customers get mildly demotivated when their data is leaked by the hackers or there’s a trial related to misuse of personal information. These companies get frowned upon.
There are interesting implications to this dichotomy.
First dichotomy. Decision makers.
In profit centres, the decisions are made by technical people, pitching ideas to the finance people.
In the cost centres, technical people are expendable and often outsourced, thus the decisions are made by people managers.
Second dichotomy. Target products.
While both groups “go to the cloud”, their needs cannot be more different from each other.
Profit centres, effectively buy compute resources. They tend to use cloud to its best, because they can fairly trivially drive the ROI of their cloud use. (Ask me how if you don’t :) )
Interestingly, some of them learn that it would be cheaper to hire a system administrator and rent a couple of racks in a data centre. (I’m sure Basecamp analysed potential cost savings while staying in the cloud, and I’d love to see their conclusions. My guess is that most of their cost is with Elasticsearch and that software, while excellent in doing what it does, is an absolute anti-cloud dinosaur from the past).Cost centres, on the other hand, are interested in SaaS products. Outsourcing is their king, because, when the financial reports are discussed, they get punished for hiring people and praised for buying services.
Again, interestingly, on one hand outsourcing office cleaning works, but on the other hand, while a bakery is happy to outsource their document management to a cloud company, outsourcing anything non-trivial (let’s say managing the VPN or DNS) is so seldom a success, I’m struggling to remember a success story. Of course many did that and got their bonuses for that, but more often than not the price was a total mess and ridiculous lead times for trivial changes. For this I have one too many examples ready in my head.
The common fate for the cost centre cloud migrations is that if the company had a data centre (private cloud in the fancy-speak), they would do a partial lift-and-shift, get stuck there, realise that they have to either double the workforce, managing their now-hybrid environment, or abandon the cloud.
The message for those doing lift-and-shift is that unless you’re planning on actually doing all the other Rs of the cloud migration, rehosting (aka lift and shift) is guaranteed to be more expensive than doing nothing.
[ Did you know? Dockerising existing applications is still lift and shift? ]
Third dichotomy. Change management and compliance.
For a moment, let’s omit POC and early day startups that often have to sacrifice quality for speed.
Compliance. For the sake of simplicity, we define it as “know that what you’re doing is keeping you on a good side of the law“.
Both of our (mature) subjects pay dear attention to these topics for, as usual, very different reasons.
Profit centres understand the value of agility. In short, agility correlates to the ability to provide business value sooner. You cannot wait a year to fill your car’s tank. Unless it’s a DeLorian, of course. Bringing value quickly and often is a good thing, but when you’re driving really fast, you better be in control of what you’re doing. Thus code reviews, automated tests, meaningful commit messages, feature flags, infrastructure as code, etc. Profit centres love change but understand the need to be in control.
Cost centre’s most popular KPI is the stability of the service they provide. Stability, without dedicated qualified personnel (everything’s outsourced to the lowest bidder, remember?), can only be implemented using the Standard Operating Procedure™. This is guaranteed to be slow - I’ve seen companies creating a change ticket for a server reboot. When it comes to the cloud, profit centres normally see it as just a bigger data centre and everything should be static (change, without ticket means trouble). In the Pets vs Cattle analogy, they are strongly on the side of pets. In many ways, this is anti-cloud.
It’s interesting, that while there’s traditionally a view that AWS and Azure are rivals, everything I’ve seen so far tells me that AWS’ target audience is profit centres and Azure’s - cost ones. First talk to the techies, second to the managers. In recent years both have tried to bring features for the other camp, but, ironically, haven’t changed the language they use with the new target audience.
Take care and stop wars.
PS. This post is 100% GPT-3 free.